feedback/adapter_meta.json

{
  "style": "feedback",
  "sample_count": 1200,
  "rank": 16,
  "base_model": "Qwen/Qwen2.5-14B-Instruct",
  "training_mode": "hf",
  "use_4bit": true
}